Separating conscious and unconscious influences of memory: Measuring recollection.

Abstract

How can conscious and unconscious influences of memory be measured? In this article, a processdissociation procedure (L. L. Jacoby, 1991) was used to separate automatic (unconscious) and consciously controlled influences within a task. For recall cued with word stems, automatic influences of memory (a) remained invariant across manipulations of attention that substantially reduced conscious recollection and (b) were highly dependent on perceptual similarity from study to test. Comparisons with results obtained through an indirect test show the advantages of the process-dissociation procedure as a means of measuring unconscious influences. The measure of recollection derived from this procedure is superior to measures gained from classic test theory and signal-detection theory. The process-dissociation procedure combines assumptions from these 2 traditional approaches to measuring memory. Dissociations between performance on direct and indirect tests of memory supply examples of effects of the past in the absence of remembering (for reviews, see Hintzman, 1990; Richardson-Klavehn & Bjork, 1988). In an indirect test, subjects are not asked to report on memory for an event as they would be in a direct test, such as in a test of recognition or recall; rather, they engage in some task that can indirectly reflect memory for the occurrence of that event. Word stemand fragment-completion tasks are among the most popular indirect tests of memory (e.g., Graf & Mandler, 1984; Tulving, Schacter & Stark, 1982; Warrington & Weiskrantz, 1974). For a stem-completion task, subjects might read the word scalp and then be presented with the stem sea— with instructions to complete that stem with the first word that comes to mind. Evidence of automatic influences of memory that are dissociated from performance on a direct test is provided by the finding that prior presentation of a word increases the likelihood of that word being used to complete a stem, even though a direct test reveals no memory for the prior presentation of the word. Some of the most striking examples of dissociations come from the performance of patients suffering a neurological deficit. Korsakoff amnesics, for example, show near-normal effects of memory in their performance of a stem-completion task, even though